Summary A WestJet Airlines Boeing 737-700 (registration C-GWJF, serial number 32766), operating as Flight 798 (WJA798), was on a scheduled passenger flight from Toronto, Ontario, to Halifax, Nova Scotia. Just before touchdown on Runway14 in low-visibility conditions, the aircraft rolled right and moved toward the right side of the runway. The aircraft then rolled to the left, and the left wing struck the runway. None of the passengers or crew members were injured, and the aircraft taxied to the terminal. The incident occurred at 1924 Atlantic standard time, during the hours of darkness. Ce rapport est galement disponible en franais. Other Factual Information The flight crew reported for duty in Regina, Saskatchewan, at 0910 Atlantic standard time1 on the morning of 25December2005 to operate a series of flights from Regina to Calgary, Toronto, and Halifax, where their duty day would end. While on the ground in Toronto, the crew requested the latest runway visual range (RVR) report for Halifax, from the operator's dispatch personnel, through the aircraft communications addressing and reporting system (ACARS). They received a reply advising that the RVR at Halifax was 1100feet, or 100feet below the minimum approach ban2 visibility of 1200feet. The report did not specify, nor did the crew ask for, the applicable runway or time of the RVR report. The RVR report was actually for Runway14, and the RVR at the time on Runway23 was 900feet. There was no guidance for dispatchers in the company operations manuals on informing flight crews of the applicable runway or time for RVR reports. The captain had requested that additional fuel be loaded on the aircraft because of the reported weather in Halifax. The captain's plan was to conduct a CategoryII instrument landing system (ILS) approach to Runway23 at Halifax, if the visibility was at or above landing limits when they arrived. St John's, Newfoundland and Labrador, had been filed as the alternate airport if a landing in Halifax was not possible. The passengers were advised of the poor weather in Halifax, and of the possibility of not being able to land. The aircraft departed Toronto at 1748with 132passengers and 6crew members on board. For this leg of the flight, the co-pilot, who was in the right seat, was the pilot flying (PF). The captain, who was in the left seat, was the pilot not flying (PNF). While en route to Halifax, the crew requested and received another RVR report for Halifax through ACARS. The report indicated that the RVR at Halifax had improved to 1400feet, or 200feet above the minimum approach ban visibility. The report was for Runway14, and again there was no indication of which runway the report referred to. The crew continued to assume that it was for Runway23. This assumption was reinforced when the crew also received the automatic terminal information service (ATIS) report for Halifax, which indicated that the active runway was Runway23. Based on the ATIS information, the crew planned for a CategoryII ILS approach for Runway23 with an auto-land, and completed an approach briefing. The aircraft was being flown on autopilot, and it remained on autopilot until just before landing. The crew requested and received descent clearance from the Moncton Area Control Centre. While in descent through 16000feet, the crew contacted Halifax terminal arrival control and informed the controller that they were planning the CategoryII ILS approach and landing on Runway23. The controller informed the crew that the RVR for the approach end of Runway23 was 1000feet, and that the RVR for Runway14 was 1200feet. A short time later, because the RVR was below limits on Runway23, the crew requested and received a clearance for the CategoryI ILS on Runway14. This required the crew to reprogram the flight management system (FMS) computer, and re-brief for the new approach. About 10 minutes before landing, the crew was advised that another aircraft had just landed on Runway14, and that the pilots of that aircraft had reported that they had the runway lights visual at 250feet above ground level (agl). This is 50feet above the decision height (DH) for the ILS approach to Runway14. When WJA798 was established on final approach, the crew contacted the Halifax control tower and was informed that the wind was calm and that the RVR for Runway 14was now 1400feet. The tower controller cleared WJA798 to land. The final approach on autopilot was stabilized and uneventful. The PNF observed the approach lights at 200feet agl (the DHfor the approach), and the PFtransitioned from monitoring the flight instruments to acquiring visual references for the landing. After passing the approach lights, the runway environment appeared very dark and the visibility was such that finding visual cues was difficult for thePF. The aircraft remained aligned with the runway and on the ILS glide path until, at 67feetagl,3 the PF disconnected the autopilot. Immediately after, the PFunintentionally rotated the control column approximately 30 degrees clockwise towards right-wing-down, and also moved it slightly aft. This resulted in the aircraft rolling to the right and levelling off. The aircraft went above the glide path, and deviated three degrees from the runway heading to a heading of 147degrees. At 22 feet agl, the PF unintentionally depressed the take-off/go-around (TOGA) switch on the throttles instead of the auto-throttle disconnect switch. This was followed by immediate selection of the auto-throttle disconnect. The operator's standard operating procedures (SOPs) require that the auto-throttles be disconnected at no lower than 50feetagl. The aircraft deviated to the right of the runway centreline. The PFattempted re-alignment using rapid and aggressive left and right control wheel movement. The captain gave verbal instructions to descend and turn left almost immediately after the deviations from the approach path, indicating that he had sufficient visual cues to correctly assess the aircraft's displacement. The aircraft touched down firmly on the left main landing gear at about 2500feet from the runway threshold, between the centreline and the right edge of the runway, with 16degrees of left bank. Concurrently, the left wing contacted the runway surface for approximately one-half second. The left main landing gear strut then extended to nearly full length, and the left bank increased to 18degrees. The wing contacted the runway again for approximately two seconds, and simultaneously, the aircraft heading deviated left to 136degrees, or 8degrees left of the runway heading. The aircraft settled onto both main landing gears five seconds after the left main gear made contact, approximately 3550feet beyond the runway threshold. After the nose gear touched down, heavy wheel braking was used to slow the aircraft. Eight seconds after nose gear touchdown, after being prompted, the PFapplied reverse thrust on both engines. Deployment of reverse thrust occurred approximately 5300feet beyond the runway threshold. The aircraft slowed to taxi speed with approximately 500 feet of runway remaining. The aircraft taxied uneventfully to the assigned gate at the terminal. The engines were shut down at the gate. During normal ground inspection of the aircraft, it was determined that the left wing tip, outboard leading edge flap, aft position light, and outboard flap canoe tail cone had been damaged. The aircraft was removed from service until repairs could be affected. There were no other discrepancies or malfunctions found with the aircraft systems. The flight data recorder (FDR) and cockpit voice recorder (CVR) were removed and sent to the TSB Engineering Laboratory for analysis. High-quality data was recovered from both recorders. The CVR circuit breaker had not been pulled after the damage was discovered, and as a result, only the final five minutes of the approach and landing were captured on the two-hour recording. The CVR continued to operate after the aircraft had landed resulting in occurrence information being overwritten; therefore, the investigation team was deprived of possible important information relative to the occurrence. There have been at least eight other occurrences where the CVR had been overwritten because of the limited duration of recording and/or the recorder had been allowed to run after the occurrence without recognition that data were being lost (see AppendixB for details). The TSB recently issued an Aviation Safety Advisory (A060008-1) with respect to cockpit voice recorder duration. In addition to increased CVR recording duration, there needs to be enhanced awareness by flight crew and maintenance personnel of the potential safety information loss if a CVRis allowed to continue to record after a significant safety event. Halifax International Airport is served by two runways: 23/05 and 14/32. Runway14 is 7700feet long by 200feet wide and has a CategoryI ILS approach with a simplified short approach lighting system with runway alignment lights. Runway23 is 8800feet long by 200feet wide and has a CategoryII ILS approach. The approach and runway lighting system on a CategoryII runway is superior to a CategoryI system in that it has an enhanced approach lighting system along with centreline and touchdown zone lighting embedded in the runway surface. These enhanced visual aids assist the pilot in visually aligning the aircraft with the runway centreline. The DH on the Category I and II approaches are 200and 100feet agl respectively. Both runways are served by RVR transmissometers, and therefore, the approaches to the runways are subject to approach ban limits. The ban prohibits pilots from conducting an approach to either a CategoryI or II serviced runway when the RVRis below 1200feet. The observed airport weather conditions at Halifax at the time of the incident were as follows: surface wind calm; ground visibility 1/8statute mile in fog; vertical visibility 100feet; temperature 1C, dew point 1C; and RVR Runways23 and14, 1000feet and 1200feet, respectively. There had been very little change in the weather for several hours before the incident. The captain had a total of 11000hours of total flying experience with 3200hours on the Boeing737. The co-pilot had approximately 6570hours of total flying experience of which 620 were on the Boeing737. The flight crew were certified and qualified for the flight in accordance with existing regulations. The following table represents the flight crew experience and workload in hours. The following is an extract from the operator's SOPs with respect to CategoryI approaches: MONITORED APPROACH GENERAL Low visibility CATI approaches are required to be flown using the autopilot and monitored approach procedures. A Monitored Approach Procedure shall be utilized during: A CATI precision approach anytime reported RVR is reported below 2600feet or, in the absence of RVR, prevailing visibility is below sm. Prior to commencing the approach the First Officer assumes control of the aircraft and the Captain will monitor the approach. To facilitate the transfer of control, the Captain will position his hand at the base of the thrust levers no later than 1000feet AGL. The Captain will assume control of the aircraft prior to landing and the First Officer will monitor the flight instruments for the duration of the approach, landing and rollout. If a go around is required after the Captain assumes control of the aircraft the Captain will fly the missed approach and the First Officer will monitor the flight instruments. In order to maintain the flow of the Monitored Approach, all callouts shall be completed at the appropriate altitudes regardless of when the runway environment is encountered. The First Officer controls the autopilot until DA(H) where the Captain calls for control if the required visual reference for landing is met. The autopilot remains coupled until at least Decision Altitude and no later than 50feet. If, at DA(H), the Captain makes no callout or calls "GO AROUND" the First Officer shall carry out a go-around. If the Captain elects to land, the First Officer will continue to monitor his flight instruments until touchdown and callout any deviations. This technique balances the cockpit workload at a critical stage of the approach (transition to landing), and improves decision making with respect to continuing for a landing or abandoning the approach. It usually also results in the more experienced flight crew member landing the aircraft when the weather conditions are at minimums. The crew did not brief for, or carry out, a monitored approach. A number of occurrences have been investigated by the TSB in which inadequate visual references during the final stages of an approach contributed to an accident. These occurrences share a number of commonalities. All were conducted during darkness with visibilities less than those recommended on the Canada Air Pilot (CAP) approach plate to runways served by a CategoryI ILS system. In these occurrences, the crew had sight of the runway environment at minimums and elected to land, but subsequently had difficulty acquiring sufficient visual references to maintain aircraft alignment with the runway. The Transportation Safety Board of Canada has previously identified safety deficiencies associated with conducting approaches in low visibilities, and has made recommendations to Transport Canada that could mitigate the deficiencies. The Safety Action section of this report identifies the current initiatives by Transport Canada to address these recommendations.